1. Field of the Invention
The present invention relates to an image processing apparatus suitable for preventing counterfeiting of copy-prohibited objects, such as paper money, securities and the like.
2. Description of the Related Art
Recently, it has been proposed to mount in a color copier an image recognition processing function or the like for preventing counterfeiting of paper money, securities and the like. That is, when image data to be copied is determined to represent an image for which counterfeiting is to be prevented, normal copy output of the image is prevented, for example, by (1) providing print output by painting the entire image with a certain color, (2) providing print output by superposing a certain symbol or figure on the image, (3) providing print output by changing the color or the image size, or (4) providing no print output.
A copier is basically designed to faithfully copy an object mounted on an original-mount, except for a copy-prohibited object, such as paper money or the like. Accordingly, if an image which resembles paper money (an image other than paper money which can be legally copied) is input, the copier must recognize that the image is not paper money, and perform a copying operation of the original image. Hence, very precise recognition accuracy is requested for determining whether or not an image to be recognized is an image for which counterfeiting is to be prevented.
Recently, the quality of an image obtained by a color scanner or a color printer, which is less expensive than a color copier, is greatly improved, and it becomes possible to perform a counterfeiting operation by connecting such an apparatus to a personal computer, i.e., by combining apparatuses other than a color copier. Accordingly, even in an image processing system using an inexpensive color scanner or color printer, countermeasures for preventing counterfeiting of paper money, securities and the like by mounting an image recognition function or the like have been desired.
Various methods for determining whether or not an original is a copy-prohibited object have been proposed. In one method, a color spectrum distribution of image data obtained by scanning an original is compared with data registered in advance in a ROM (read-only memory), based on each pixel value (a set of R, G and B values corresponding to three color signals) in the image data (for example, by obtaining a sum of the absolute values of difference values between respective pixel values and stored data, or calculating a cross-correlation value between distributions). In another method, an evaluation value is calculated by comparing a synthetic image pattern of a part or the entirety of an original with pattern data registered in advance in a ROM or the like.
Recently, it became possible to incorporate information indicating a copy-prohibited object within an original, serving as a printed matter, using a technique called “digital watermark That is, when printing an original, serving as a copy-prohibited object, a printed matter is obtained from image data (electronic image information) including copy-prohibited-object information generated by embedding second digital information (sub-information) indicating a copy-prohibited object in image data (electronic image information), serving as original first digital information (main information). This copy-prohibited-object information (sub-information) can be extracted from the image data including the copy-prohibited-object information before printing the original. Furthermore, if consistency between conditions when forming the printed matter from the image data (electronic image information) including the copy-prohibited-object information and conditions in reading and extraction processing is obtained, it is also possible to extract the copy-prohibited-object information from electronic image data obtained by reading the printed matter formed from electronic image information including the copy-prohibited-object information by an original-reading device, such as a color scanner or the like. In consideration of these features, application of the technique called “digital watermark” to prevent illegal copying of an original, serving as a printed matter, has been attracting notice.
However, as described above, in order to extract digital-watermark data embedded using the digital watermark technique, it is necessary to obtain consistency between conditions in reading and extraction and how watermark information is embedded in original-image data (i.e., at what position and with what arrangement in the original-image data the watermark information is embedded).
Accordingly, when reading a printed matter (an original), formed from electronic image information having embedded digital-watermark information, by an original-reading device, such as a color scanner or the like, the direction the original is mounted on the surface of an original-mount greatly influences the difficulty in extraction of digital-watermark information.
When recognizing a copy-prohibited object by comparing a color spectrum distribution of image data obtained by scanning an original with data registered in advance in a ROM, or by calculating an evaluation value by comparing a synthetic image pattern of a part or the entirety of the original with pattern data registered in advance in a ROM or the like, if the direction (angle) or the position of the mounted original is not fixed, the total amount of calculation operations for comparison with spectrum data registered in advance or image-pattern data tends to greatly increase in order to deal with various directions and positions.
In an inexpensive color scanner or color printer, it is desirable to prevent counterfeiting of a copy-prohibited object with a cost lower than in the case of using a relatively expensive color copier. For that purpose, it is preferable to recognize a copy-prohibited object by mainly using software instead of mainly using hardware having a large amount of electronic circuits and the like.
On the other hand, when determining a copy-prohibited image requiring precise recognition accuracy mainly by software, the total amount of calculation operations is large, and the processing time required for the recognition and determination tends to be greatly increased.
That is, if the amount of calculation operations for the recognition and determination is reduced, it generally tends to be difficult to determine a copy-prohibited image requiring precise recognition accuracy such that, for example, it becomes difficult to detect a copy-prohibited object, such as paper money or the like, read by intentionally changing reading conditions (such as a mounting angle or a mounting position of an original on a reading surface), or that when an image resembling paper money (an image other than paper money which is legally allowed to be copied) is input, the image is erroneously recognized as a copy-prohibited image and normal image output is not performed, thereby impairing the original function of an image processing system.